请求提供SSv2上训练日志

[yinshicheng]

嗨！亲爱的作者，请问您是否有保留您在ssv2上微调IN1k模型的训练日志，我现在在尝试复现您论文中报告的成绩以进行进一步研究，但是模型的前期表现不是很好，我担心是否是出了什么问题，希望能得到您的训练日志以对比，下面是可视化的训练日志，其中蓝色的是在1/10SSv2上训练的结果，黄色是在完整SSv2上训练的，它看起来甚至不如小号数据集的表现，我已经检查了我的完整数据集，与预期的文件数量一致。

日志文件如下：

{"train_lr": 4.087579617834395e-05, "train_min_lr": 3.076081301129866e-08, "train_loss": 4.954533441759041, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 4.352651517207806, "val_acc1": 10.388247894599946, "val_acc5": 28.997498573465183, "epoch": 0, "n_parameters": 7004239}
{"train_lr": 0.000120687898089172, "train_min_lr": 9.08228881866906e-08, "train_loss": 4.491472584085992, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 3.4437925283725446, "val_acc1": 24.28767516582408, "val_acc5": 50.39551334995384, "epoch": 1, "n_parameters": 7004239}
{"train_lr": 0.00020050000000000005, "train_min_lr": 1.5088496336208257e-07, "train_loss": 4.17860123982838, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 3.00919756430846, "val_acc1": 31.1808871765654, "val_acc5": 59.98466519903777, "epoch": 2, "n_parameters": 7004239}
{"train_lr": 0.00028031210191082804, "train_min_lr": 2.109470385374746e-07, "train_loss": 4.0046276775971785, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.684146716961494, "val_acc1": 36.84720434893495, "val_acc5": 66.65187063270278, "epoch": 3, "n_parameters": 7004239}
{"train_lr": 0.00036012420382165603, "train_min_lr": 2.7100911371286653e-07, "train_loss": 3.913576457697483, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.5948487318479097, "val_acc1": 38.832836932290775, "val_acc5": 67.93123147908899, "epoch": 4, "n_parameters": 7004239}
{"train_lr": 0.0003996359870051319, "train_min_lr": 3.0074344766802523e-07, "train_loss": 3.827232006375224, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.4842891381337093, "val_acc1": 41.38348661356991, "val_acc5": 71.13568733025043, "epoch": 5, "n_parameters": 7004239}
{"train_lr": 0.0003974550400281364, "train_min_lr": 2.9910219028788265e-07, "train_loss": 3.7539557550240863, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.376421308058959, "val_acc1": 42.50948546382896, "val_acc5": 72.41101245170394, "epoch": 6, "n_parameters": 7004239}
{"train_lr": 0.00039311621241710056, "train_min_lr": 2.95837033953092e-07, "train_loss": 3.7056594081138283, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.309735374725782, "val_acc1": 44.192429974462286, "val_acc5": 73.90830805580902, "epoch": 7, "n_parameters": 7004239}
{"train_lr": 0.0003866670412752908, "train_min_lr": 2.909837523997355e-07, "train_loss": 3.6516650444273697, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.2990021421359135, "val_acc1": 44.430544536755924, "val_acc5": 74.55807830369619, "epoch": 8, "n_parameters": 7004239}
{"train_lr": 0.000378178185071309, "train_min_lr": 2.845955191961262e-07, "train_loss": 3.6162801605901365, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.250480229120988, "val_acc1": 44.98749006903964, "val_acc5": 75.42578349158497, "epoch": 9, "n_parameters": 7004239}
{"train_lr": 0.0003677426494901356, "train_min_lr": 2.767423251620669e-07, "train_loss": 3.572421827128297, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.180074443266942, "val_acc1": 47.15069945136247, "val_acc5": 76.47510136594995, "epoch": 10, "n_parameters": 7004239}
{"train_lr": 0.0003554747684433739, "train_min_lr": 2.6751021153478e-07, "train_loss": 3.529850111665646, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.197427235199855, "val_acc1": 47.380742226321296, "val_acc5": 76.43877881683096, "epoch": 11, "n_parameters": 7004239}
{"train_lr": 0.0003415089514029928, "train_min_lr": 2.570003272831112e-07, "train_loss": 3.495167897993005, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.0958071204332205, "val_acc1": 48.89014578362106, "val_acc5": 78.05715012987201, "epoch": 12, "n_parameters": 7004239}
{"train_lr": 0.00032599821078301895, "train_min_lr": 2.4532782089825575e-07, "train_loss": 3.453393871898087, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.1253006838835202, "val_acc1": 48.10719317914172, "val_acc5": 77.84728624068447, "epoch": 13, "n_parameters": 7004239}
{"train_lr": 0.0003091124855035178, "train_min_lr": 2.3262057880279557e-07, "train_loss": 3.426993333723981, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.058064203995925, "val_acc1": 49.61256069128234, "val_acc5": 78.76342183235569, "epoch": 14, "n_parameters": 7004239}
{"train_lr": 0.0002910367791042166, "train_min_lr": 2.1901782420029013e-07, "train_loss": 3.393374560817432, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.023080690090473, "val_acc1": 50.395513399219446, "val_acc5": 79.35265430459985, "epoch": 15, "n_parameters": 7004239}
{"train_lr": 0.0002719691328070099, "train_min_lr": 2.0466859171672372e-07, "train_loss": 3.3712447294238124, "train_loss_scale": 0.0, "train_weight_decay": 0.09999999999999756, "val_loss": 2.027491133029644, "val_acc1": 50.20986479074809, "val_acc5": 79.48180090177337, "epoch": 16, "n_parameters": 7004239}